Thomson Reuters’ deal to purchase Casetext has driven the legal tech hype cycle to a fever pitch. Don’t get me wrong. I am a big fan of Casetext. I have been an admirer for over a decade. More than once, I have watched them not just beat the market, but redirect the market and invent whole new categories of legal research products. I have spent a lot of time over the past few years musing about innovation, new product categories and market advantage.
When Casetext created the “brief analysis” tool CARA, it was three years before even one of the largest legal information companies launched a brief analyzer and it was four years before all three,; Thomson Reuters, LexisNexis and Bloomberg Law, had a brief analyzer on the market.
Note: vLex (Fastcase) launched the Vincent “brief checker” in the international market in 2018., two years after the launch of CARA. The above graphic was focused on the U. S. market.
Similarly, with Compose, Casetext introduced “parallel search” and a new category of concept searching was born. Or as Casetext co-founder Pablo Arredondo likes to exclaim “parallel search freed lawyers from the prison of the keyword.” This time The market responded in less than two years. LexisNexis launched “Fact and issue Finder” which leveraged extractive AI technology built on a highly tailored version of Google’s BERT to present insights to researchers. Westlaw responded with Westlaw Precision built with a large editorial team to help with machine learning. The response time in the market is growing shorter.
Industry Insiders’ Perspectives I have spoken to several legal tech industry insiders and the consensus seems to be that within 6 to 12 months, Thomson Reuters’ competitors Lexis, Bloomberg Law, and vLex (formerly Fastcase) are likely to have developed capabilities which can compete with CoCounsel. No one is starting from scratch. According to Ed Walters, Chief Strategy Officer at vLex, “we already have global AI products in the labs. We don’t release vaporware, but these products are coming no matter who owns Casetext.”
I reached out to Jeff Pfeifer, Lexis’ Chief Product Officer, Canada, Ireland, UK and USA to get a sense of their AI development timeline. “LexisNexis is confident that our current development plans for Lexis+ AI will deliver a solution that meets or exceeds capabilities today from other legal generative AI solutions. We remain on track to delivery our solution later this summer and we continue to benefit from the valuable feedback of law firms participating in our Commercial Preview. LexisNexis has worked to develop a deep bench of AI talent in the last five years and our new Lexis+ AI development benefits from the work we have done over this time period. To date, we’ve released more than 15 feature capabilities leveraging extractive AI, including the newly released Agreement Analysis on Lexis+.”
Calculating the Cost of Market Advantage?
Let’s face it, the announcement of TR’s purchase of Casetext was a “hair on fire moment” for their competitors, although rumors of the deal had been circulating for months. Casetext launched CoCounsel at Legal Tech in January – so everyone knew where the market was heading. CoCounsel currently offers 8 skills – document review, deposition preparation, searching a database, legal research memo drafting, summarization, contract analysis, contract policy compliance, and market check.
Regulatory Approval The Thomson Reuters – Casetext deal won’t be finalized until it survives regulatory approval. Let’s assume that it could take six months, but given Thomson Reuters’ position in the marketplace maybe it will take longer, say, 9 months?
Integration I assume the integration of Casetext into Thomson Reuters’ products will take time as well. Six months? A year? An executive from a company acquired by TR (who chooses to remain anonymous) said that being merged into TR was like being ingested by a machine. The uniqueness of the acquired company is squeezed into the parameters of Thomson Reuters’ business model. This does not happen overnight.
The Competition Meanwhile TR’s competitors, Bloomberg Law, LexisNexis, Wolters Kluwer and vLex presumably need no government approvals to continue working internally on AI products.
Hypothetically then, while the benefits of the Casetext acquisition are on hold, TR’s competitors are free to develop competing products.. At the earliest TR may not be able to integrate CoCounsel technology into their products until June 2024.
The Cost of Time – $216M per month of lead time? So let’s assume that after the regulatory approval and integration, Thomson Reuters ends up with a 3 months market advantage compared to their competitors. At $650M, Thomson Reuters is paying about $216,666.667.00 per month to “maybe” be ahead of their competitors. If they have 12 months of market advantage that is a mere $5,416,666 .00 per month. But the timelines for the TR integration of CoCounsel technology and the launch of competitor products are all speculation.
One thing is for sure.. Who will pay the cost at the end of the day? You, the customers!
Market Assessment
Given the pace of technological change, It won’t take too long before we understand the whether this is a good deal for Thomson Reuters. It certainly looks like a good deal for Casetext investors. I hope it is a good deal for Casetext co-founders Jake Heller, Pablo Arredondo and Laura Safdie who repeatedly reenergized the legal market with their groundbreaking creativity and reimagining of legal research.
Did Thomson Reuters simply buy the Casetext suite of software including CoCounsel? Are they also getting the Casetext “brain trust”? Could the wild sparks of Silicon Valley creatives still fly inside the cavernous TR corporate culture?
Is it a good deal for the market? In the short term – it may speed Generative AI launches from competitors. Competition may tamp down potential price increases. But in the long term it is likely that the cost of Casetext will be shifted to Thomson Reuters’ existing customers across various product lines even if they are not getting any immediate benefit from the new generative AI capabilities.